Software
cbi_partitions
cbi_partitions is a Python library implementing Conformalized Bayesian Inference (CBI) for clustering problems based on partition-valued Monte Carlo posterior samples. It provides point estimation of the data clustering, construction of a credible set of partitions with guaranteed posterior coverage using conformal prediction principles, and density-based clustering to explore posterior multimodality. The method is introduced in the paper “Conformalized Bayesian Inference, with Applications to Random Partition Models”.